Sudden cardiac arrest is a leading cause of death and disability in the U.S. Even if resuscitation achieves return of spontaneous circulation (ROSC), >50% of patients suffer fatal or crippling brain injury. Recent data suggest that exposure to hyperoxia (i.e. supranormal arterial oxygen partial pressure due to an excessively high fraction of inspired oxygen [FiO2]) after ROSC may accelerate oxidative stress, worsen reperfusion injury in the brain, and contribute to poor outcome. Prolonged hyperoxia exposure is extremely common in post- resuscitation care; however, no clinical study has tested if rapidly reducing FiO2 after resuscitation reduces oxidative stress and improves outcomes. Our hypothesis is that rapid FiO2 de-escalation after ROSC reduces oxidative stress and is associated with improved neurological, neuropsychological, and functional outcomes. To test this hypothesis, we will pursue the following Specific Aims using an experienced six center Emergency Department (ED)-based research consortium (Emergency Medicine Shock Research Network). Our first aim is to test if a protocol for rapid FiO2 de-escalation after ROSC lowers in vivo oxidative stress during post- resuscitation care. We will conduct a multicenter prospective before-and-after protocol implementation study testing the effects of a protocol for rapid FiO2 de-escalation after ROSC among 266 adult ED patients resuscitated from out-of-hospital ventricular fibrillation cardiac arrest. All sites currently employ uniform methods of pos-resuscitation care including therapeutic hypothermia. In the before phase, patients will receive standard care. In the after phase, patients will receive standard care plus rapid FiO2 de-escalation after ROSC via implementation of a clinical protocol. We will compare the most sensitive and specific in vivo biomarkers of oxidative stress - plasma and urine isoprostanes and isofurans measured by mass spectroscopy - over the first 24 hours between groups. Our second aim is to test if the rapid FiO2 de- escalation protocol is associated with reduced neurological disability. We will determine the Modified Rankin Scale (mRS) (a well-validated scale of neurological disability ranging from 0=no disability/asymptomatic to 6=death) at hospital discharge, and compare the proportion of patients with good neurological outcome (defined as a mRS <3) between groups. Our third aim is to test if the rapid FiO2 de-escalation protocol is associated with improved neuropsychological and functional outcomes. We will compare neuropsychological and functional outcomes between groups among survivors at 90 days using validated instruments across five cognitive domains (memory, executive function, lexical-semantic, visuoperceptual, psychomotor) and ability to return to work. Significance: This project will generate new, critically important knowledge about a fundamental element of post-cardiac arrest care. Research on new methods to improve outcomes from this devastating condition is a crucial and urgent priority. Given the magnitude of death and disability that cardiac arrest imparts upon Americans, even modest improvements would provide a major public health benefit.